Characterization and Estimation of Heavy Rainfall Using Dual-polarization Radar Observations During 21 July 2012 Beijing Flood Event

As one of the most hazardous natural disasters in both rural and urban areas, flooding accounts for about one-third of all geophysical hazards over the world (Adhikari et al. 2010). Unfortunately, the rapid urbanization and land degradation have made the densely populated areas even more vulnerable to flash floods, especially in regions where the authorities have little presence and the resources are lacking to ensure basic social services, policing, running water, and sewerage, etc.

On 21 July 2012, the capital of China, Beijing, and its surrounding areas experienced extreme rainfall and flooding. The severe storm lasted for about 16 hours and the peak storm total reached as high as 300 mm. It was reported as the heaviest storm event since 1951. The flooding caused 79 fatalities and around 1.6 billion USD in damages. Therefore, high spatiotemporal-resolution rainfall monitoring in such region is critical in terms of human safety and property. However, due to the damages, power loss, and communication interrupting, some of the Automatic Weather Stations (AWS) were not be continuously working during the severe storms, which resulted in un-reliable rainfall observations. To this end, weather radars are often used for rainfall estimation because of the capability of conducting spatially continuous observations over a large area with small temporal sampling intervals (e.g., Bringi and Chandrasekar 2001; Cifelli and Chandrasekar 2010; Chen and Chandrasekar 2015).

In this paper, the quality control of AWS data will be presented based on the measurements collected during the 21 July 2012 Beijing flood event. In addition, we characterize the signatures of this heavy rainfall event using dual-polarization observations from a C-band radar that is operated by Beijing Meteorological Bureau (BMB). In particular, the distribution of dual-polarization observables will be investigated. Radar rainfall products are produced using Z-R relation, as well as dual-polarization radar rainfall algorithms. It has been shown that dual-pol rainfall products are superior to single-pol products based on the evaluation using quality controlled AWS data.